IoT based Novel Face Detection Scheme using Machine Learning Scheme
K. Sumathi, D. Vishnu Sakthi, G. Nirmala, P. Sellamuthu, Ranjan Walia, Mohd Usman · 2022 International Conference on Advances in Computing, Communication and Applied Informatics (ACCAI) · 2022
Face detection scheme is the most significant operation of the Digital Image Processing domain, in which it is supportive to many industries to resolve the security issues and the associated time savvy functionalities. At the moment, safety concerns about the Internet of Things (IoT) are widespread. Facial recognition in the presence of selective concealment has developed into a critical issue impacting social protection. This research proposes a unique paradigm for face recognition by using the Internet-of-Things security environment, in which it is utilized to identify certain criminal actions. Our developed framework incorporates color and contour information into a learning algorithm and its advantage in detecting faces with extreme occlusion has been proved to be persistent. These services are divided into three categories: initially to introduce a novel face identification approach depending on the efficiency feature; next, employ a Support Vector Machine (SVM) model to generate feature representations for occluded faces. Finally, an unique sparse classification framework with a deep learning approach is created to determine whether the observed face is covered. Statistical analysis demonstrates that this approach outperforms the advancements in terms of effectiveness and resilience. At a frame rate of ten frames-per-second, the created face recognition method achieves 98.5% accuracy despite the presence of various forms of extreme deformations in faces and the developed occlusion verification technique achieves 96.5% accuracy.